Auto Emotions


If a Lamborghini Gallardo pulls up next to you at a stoplight, you're likely to feel a range of emotions (beyond, you know, envy). Its form is likely to project a sense of power, speed, and sophistication. And, ignoring the logo and other telltale details like headlight shape, bumpers, and corners, the form of the vehicle should shout out "I'm a Lamborghini!"

Auto designers, sketching with pencil and paper, shaping in clay, or smoothing lines in CAD, are out to show all that. But however well-trained they may be, the curves and contours they come up with often need only satisfy a handful of people within their company before they are slated for production. Is there a way to calculate the potential consumer's emotions when seeing the overall shape of a car?

"What do consumers see in highly aesthetic designed objects like cars?" asks Levent Burak Kara, a professor of mechanical engineering at Carnegie Mellon University. "There's not that much emphasis on the perception. It's complex because the bulk form is integrated with all the surface details. On a BMW there are lots of surface cues that give away that it's a BMW, but until today most characteristics have focused on superficial features, things you can grab off a car and hold in your hand."

Nine stages of geometric modeling of the Mustang. Image:

"Abstraction" Adaptations

Kara and his colleagues have come up with a computer graphics technique that allows them to methodically create a series of shapes that are generalizations of a car's larger form. They call the tool "Abstraction." The images it creates range from the gentle approximation of a toy car to bloated, near unrecognizable adaptations. "We take a CAD model and almost reverse engineer it until it's an extremely simple geometric file," he says. "Then we create a spectrum of models with different levels of details, of volumetric bulk details."

With the images, Kara's team has been able to determine just how far a model can deviate from an original and still maintain its brand recognition. A Mustang, for instance, still looks like a Mustang when distorted to level three or four. Beyond that, Kara's guinea pigs could no longer tell the make of the car. They've also been able to quantify just how much the larger form contributes to an auto-ogler's sense of speed, power, sophistication, or utility. It turns out that not all potential emotions can be found in the same level of abstraction. To elicit a feeling of sophistication, for instance, requires a greater level of detail than power.

Until now, designers could not change the form of an existing design, in three dimensions, in this sort of overall abstracted way. "It doesn't work from a single topology," says Kara. "What you see with each abstraction is a different volume. Before, it was points on a line, a closed network of ten curves. You would change control points on curve, but you were limited to those curves."

Conciliating Consumers

The tool will help designers and automotive executives satisfy seemingly contradictory consumer demands. Each car model has to be different, has to seem to improve and change with the latest fashion while maintaining brand identity all the while. Many of those sought-after emotions come from brand recognition alone. "BMW screwed up their design for some time, and Audi caught up with them," says Kara. "It's important that you preserve your brand."

With or without Abstraction on their computers, auto designers and car companies—and even designers from other walks of engineering—can take a single important lesson home from Kara's research. "They should get feedback from consumers from the beginning of the design cycle. Even from a rough sketch or clay model. What you establish at that point is going to be so critical," he says. "They shouldn't wait until it's all designed."

Michael Abrams is an independent writer.

Until today most characteristics have focused on superficial features, things you can grab off a car and hold in your hand.

Prof. Levent Burak Kara, Carnegie Mellon University


August 2013

by Michael Abrams,